High Player Count Raids?

[Delin]

Yea I mean I am not sure whats in common with all the zones, its generally playable at 24 its once we push past that 24 up towards 30 that the zone stability becomes an issue. I would imagine the guys in Hooptie's Haus could backup my statement on this purely because I think they have all lagged out / crashed when we entered HoH not realizing they were there as well after putting up to 40+ in zone.

Im not sure what I can do to help pinpoint the issue. I think I have done everything possible client side to reduce problems but again I am open to any suggestions that I may have overlooked. Currently I have ServerFilter=1 in eqclient.ini, everything except chats set to Hide for in game filtering, and upwards to 50 messages filtered through MQ2 so they dont constantly spam all clients. We have probably 8 auras running at the moment. I have stopped using the healing automation because I just assumed maybe reducing the amount of scripts that are run in the zone may help, but I only had one and it hasn't seemed to make much impact.

[Vaion]

Okay I made a few small adjustments. Essentially I reviewed the things the server was calling for during the "refresh" of data every few seconds. There were a few spots where the server was pulling information from the database which can cause some pretty high strain on the zone when it occurs.

I didn't move everything around to new spots, but i did highlight them/take notes and will be moving things bit by bit to try and make things better.

For this first adjustment, I moved aura's to only verify data on the initial zone in. Let me know how it pans out after next reboot.

[Vaion]

Okay I made a few small adjustments. Essentially I reviewed the things the server was calling for during the "refresh" of data every few seconds. There were a few spots where the server was pulling information from the database which can cause some pretty high strain on the zone when it occurs.

I didn't move everything around to new spots, but i did highlight them/take notes and will be moving things bit by bit to try and make things better.

For this first adjustment, I moved aura's to only verify data on the initial zone in. Let me know how it pans out after next reboot.

Just for clarification- Auras will react the same way as they used to, but before it was "checking" the aura data in the database every few seconds. Now it only checks that once upon zoning into a zone and if an aura is enabled, it will set the refresh timer as normal.

[Delin]

Went through Vallon Zek instance with 24 characters and still received some pretty hefty desync. Mobs basically get killed to 0% and bounce back up to 50% repeat with random % until everything catches up and the mob really dies.

[Vaion]

Copy. Made another round this morning that I will put in once i get time to sit and reboot the server again.

Code:
o Reduced the distance that clients receive position updates from others (client/npc) by 50% (May be adjusted based on feedback) - idea is to cut out useless position updates that are too far away for you to see anyway.
o Fixed a crash that could occur with FD when trying to wait the 2 minutes to feign off all aggro. - discovered this while snooping through the code.

[Denizen]

Great to have you tweaking things.

Wondering if events like the Zek fights actually monitor the position and update clients of everything in Drundar although irrelevant to the fight. Seems like with click ins, everything outside of the click in zone should be irrelevant to the players in the event.

[Vaion]

It definitely shouldn’t monitor anything outside of the instance zone. There are a few guesses that I have outisde of this stuff that I am probably going to set up some “stress” test environments for. My thinking is to make some instances that spawn specific events so we can narrow down to if the lag originates from scripts, client, npc or other sources

Before we get to that point tho I want to eliminate any strains I can identify in the basic code first

[Delin]

Went through tactics Tallon/Vallon instances with 24 characters. There is still definitely some desync (mob hp bouncing, delayed spell recasts, etc) however it has improved greatly from what it was (it seems on initial test). Was able to do the clear without any crippling desync to the point of lagging out so horribly that I die. However this was with just me controlling my wifes characters as well and I am a bit more sluggish than if we were both playing so there may have been less going on. Will try again later as well when she is available to see how it holds up.

Trying to use the vallon/tallon instances purely because it seems to be easy for me to replicate here and I assume this zone has much fewer scripts running than say HoH/Nadox which presumably makes it easier to pinpoint issues. Will update when I get the chance to run other stuff to see how it goes.

As a side note I would like to say that I really appreciate you putting in the time to try and fix this, since I know it probably cuts into development time for new stuff which I am excited for as well. I really enjoy the server and the encounters that you have all built here and honestly the desync is the only problem I have had.

[Delin]

Went into Rallos Zek with 24 pretty hefty desync, a lot of delay on spells / hp bouncing. He ended up despawning at 6% as a result of the desync. Was able to watch him despawn on one character that wasnt quite as desynch'd then moved to my tank which was pretty heavily desync'd and watched the fight continue for another 30 seconds.

[Denizen]

I gave up on Rallos for 24 a couple of years ago... lag cuts down your dps so much that it is better to go with a hard fast two group take down.